Update app.py

app.py

@@ -1,20 +1,122 @@
-import gradio as gr
-
-def show_instructions():
-    return (
-        "This Space hosts a command-line k-mer analysis script.\n\n"
-        "Run it locally like this:\n"
-        "python kmer_unique.py --group-dirs path/to/groupA path/to/groupB "
-        "--k-min 1 --k-max 50 --min-freq 5 --outdir kmer_results\n"
-    )
-
-demo = gr.Interface(
-    fn=show_instructions,
-    inputs=[],
-    outputs="text",
-    title="Unique k-mer Analysis (CLI tool)",
-    description="This Space is a Gradio wrapper so the repository builds. The main tool is kmer_unique.py."
-)
-
-if __name__ == "__main__":
-    demo.launch()
+import os
+import sys
+import shutil
+import uuid
+import zipfile
+
+import gradio as gr
+
+# Ensure repo root is importable on Spaces
+sys.path.append(os.path.dirname(__file__))
+
+import kmer_predict  # must be in repo root
+
+
+PERSIST_BASE = "/tmp/kmer_predict_runs"
+FASTA_EXTS = (".fa", ".fasta", ".fas", ".fna")
+
+
+def _zip_dir(folder: str, zip_path: str) -> None:
+    with zipfile.ZipFile(zip_path, "w", zipfile.ZIP_DEFLATED) as z:
+        for root, _, files in os.walk(folder):
+            for fn in files:
+                full = os.path.join(root, fn)
+                rel = os.path.relpath(full, folder)
+                z.write(full, rel)
+
+
+def run_prediction(unknown_files, kmer_zip, seqtype, mode, identity, coverage, fdr):
+    if not unknown_files:
+        raise gr.Error("Please upload at least one unknown FASTA file.")
+    if not kmer_zip:
+        raise gr.Error("Please upload the k-mer results ZIP from Space 1.")
+
+    os.makedirs(PERSIST_BASE, exist_ok=True)
+    run_id = uuid.uuid4().hex[:10]
+    run_dir = os.path.join(PERSIST_BASE, f"run_{run_id}")
+    os.makedirs(run_dir, exist_ok=True)
+
+    unknown_dir = os.path.join(run_dir, "unknown")
+    outdir = os.path.join(run_dir, "predictions")
+    os.makedirs(unknown_dir, exist_ok=True)
+    os.makedirs(outdir, exist_ok=True)
+
+    # Copy unknown FASTAs
+    for idx, f in enumerate(unknown_files, start=1):
+        src = getattr(f, "path", None) or getattr(f, "name", None) or str(f)
+        orig = getattr(f, "orig_name", None) or getattr(f, "filename", None) or os.path.basename(src)
+
+        # Ensure a valid fasta extension for readability (parser doesn't require it, but it's cleaner)
+        if not orig.lower().endswith(FASTA_EXTS):
+            orig = f"unknown_{idx}.fasta"
+
+        shutil.copy(src, os.path.join(unknown_dir, os.path.basename(orig)))
+
+    # K-mer ZIP path (ZIP-only)
+    kmer_zip_path = getattr(kmer_zip, "path", None) or getattr(kmer_zip, "name", None) or str(kmer_zip)
+    if not str(kmer_zip_path).lower().endswith(".zip"):
+        raise gr.Error("K-mer input must be a .zip file from Space 1.")
+
+    # Run prediction
+    kmer_predict.predict(
+        unknown=unknown_dir,
+        kmer_input=kmer_zip_path,
+        output_dir=outdir,
+        seqtype=seqtype,
+        mode=mode,
+        identity_threshold=float(identity),
+        min_coverage=float(coverage),
+        fdr_alpha=float(fdr),
+        group_regex=kmer_predict.DEFAULT_GROUP_REGEX,
+    )
+
+    plot_path = os.path.join(outdir, "predicted_results_summary.png")
+    csv_path = os.path.join(outdir, "predictions_by_alignment.csv")
+
+    zip_path = os.path.join(run_dir, "prediction_outputs.zip")
+    _zip_dir(outdir, zip_path)
+
+    return plot_path, csv_path, zip_path
+
+
+with gr.Blocks() as demo:
+    gr.Markdown("# K-mer Sequence Predictor")
+    gr.Markdown(
+        "Upload unknown FASTA sequences and the **kmer_results.zip** produced by the Unique k-mer Space."
+    )
+
+    unknown_files = gr.File(
+        label="Unknown FASTA files",
+        file_count="multiple",
+        file_types=[".fa", ".fasta", ".fas", ".fna"],
+    )
+
+    kmer_zip = gr.File(
+        label="kmer_results.zip (from Space 1)",
+        file_count="single",
+        file_types=[".zip"],
+    )
+
+    with gr.Row():
+        seqtype = gr.Radio(["dna", "protein"], value="dna", label="Sequence type")
+        mode = gr.Radio(["fast", "full"], value="fast", label="Mode")
+
+    with gr.Row():
+        identity = gr.Number(value=0.90, precision=2, label="Identity (full mode)")
+        coverage = gr.Number(value=0.80, precision=2, label="Coverage (full mode)")
+        fdr = gr.Number(value=0.05, precision=3, label="FDR alpha (full mode)")
+
+    run_btn = gr.Button("Run prediction")
+
+    out_plot = gr.Image(label="Prediction summary plot")
+    out_csv = gr.File(label="Predictions CSV")
+    out_zip = gr.File(label="Download all outputs (ZIP)")
+
+    run_btn.click(
+        fn=run_prediction,
+        inputs=[unknown_files, kmer_zip, seqtype, mode, identity, coverage, fdr],
+        outputs=[out_plot, out_csv, out_zip],
+    )
+
+if __name__ == "__main__":
+    demo.launch()